RESEARCH Open Access

Prophylactic inhibition of NF-κB expressionin microglia leads to attenuation of hypoxicischemic injury of the immature brainNahla Zaghloul1* , Dalibor Kurepa2, Mohammad Y. Bader1, Nadia Nagy1 and Mohamed N. Ahmed1

Abstract

Background: Periventricular leukomalacia (PVL), a devastating brain injury affecting premature infants, is the mostcommon cause of cerebral palsy. PVL is caused by hypoxia ischemia (HI) and is characterized by white matternecrotic lesions, microglial activation, upregulation of NF-κB, and neuronal death. The microglia is the main cellinvolved in PVL pathogenesis. The goal of this study was to investigate the role of microglial NF-κB activity and itsprophylactic inhibition in a neonate mouse model of HI.

Methods: Transgenic mice with specific knockout NF-κB in microglia and colony stimulating factor 1 receptor Crewith floxed IKKβ (CSF-1R Cre + IKKβflox/wt ) were used. Postnatal day 5 (P5) mice underwent sham or bilateraltemporary carotid artery ligation followed by hypoxia. After HI insult, inflammatory cytokines, volumetric MRI,histopathology, and immunohistochemistry for oligodendroglia and microglial activation markers were analyzed.Long-term neurobehavioral assessment, including grip strength, rotarod, and open field testing, was performed atP60.

Results: We demonstrate that selective inhibition of NF-κB in microglia decreases HI-induced brain injury bydecreasing microglial activation, proinflammatory cytokines, and nitrative stress. Rescue of oligodendroglia isevidenced by immunohistochemistry, decreased ventriculomegaly on MRI, and histopathology. This selectiveinhibition leads to attenuation of paresis, incoordination, and improved grip strength, gait, and locomotion.

Conclusion: We conclude that NF-κb activation in microglia plays a major role in the pathogenesis of hypoxicischemic injury of the immature brain, and its prophylactic inhibition offers significant neuroprotection. Using aspecific inhibitor of microglial NF-κB may offer a new prophylactic or therapeutic alternative in preterm infantsaffected by HI and possibly other neurological diseases in which microglial activation plays a role.

Keywords: Periventricular leukomalacia, White matter brain injury, Cerebral palsy, Hypoxia ischemia, NF-κB,Microglia, Oligodendroglia, Neuroprotection

© The Author(s). 2020 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License,which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you giveappropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate ifchanges were made. The images or other third party material in this article are included in the article's Creative Commonslicence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commonslicence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtainpermission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to thedata made available in this article, unless otherwise stated in a credit line to the data.

* Correspondence: nzaghloul@peds.arizona.edu1Department of Pediatrics, Division of Neonatology, University of Arizona,1501 N. Campbell Avenue, Tucson, AZ, USAFull list of author information is available at the end of the article

Zaghloul et al. Journal of Neuroinflammation (2020) 17:365 https://doi.org/10.1186/s12974-020-02031-9

BackgroundPeriventricular leukomalacia (PVL) is a major neuro-pathologic white matter brain injury and the most com-mon cause of cerebral palsy (CP) in premature infants.In the USA, about 65,000 very low birth weight infants(< 1500 g) are born annually. Ten percent of those in-fants show signs of CP and 25–50% display cognitive orbehavioral deficits [1]. In extremely low birth weightinfants (birth weight < 1000 g), the incidence of CP is es-timated to be approximately 20% [1]. Risk factors impli-cated in the development of PVL include prematurityassociated with immature cerebrovascular development,hypoxic ischemic insults (HI) with lack of appropriateautoregulation of cerebral blood flow, free radical pro-duction, energy deprivation, and chorioamnionitis. Af-fected infants show definitive signs of cerebral palsysuch as spastic diplegia or seizures, mental retardation,visual and hearing impairment, scoliosis, or incontinenceby 6–9 months of age.PVL pathology is characterized by focal necrosis,

microgliosis, and a decrease in pre-myelinating oligo-dendrocyte precursors, along with an arrest of differenti-ation, leading to hypomyelination and ventriculomegaly.Candidate pathological mechanisms that drive oligo-dendroglial death and white matter injury include oxida-tive stress, excitotoxicity, and inflammation [2–4]. Inneonate rat models of PVL, microglia exhibit a robustand transient response to white matter injury [5]. Micro-glial density increases rapidly within 24 h after hypoxiaischemia (HI) and continues to rise until 96 h post-HI.Since microglial activation coincides with white mattercell death following HI, we sought to determine whetherinhibiting microglial activation is neuroprotective in aneonate mouse model of PVL.The classical NF-κB signaling is a major regulator of

microglial inflammation. Additionally, NF-κB regulatesgene expression of proinflammatory cytokines, chemo-kines, enzymes, and adhesion molecules, many of whichare upregulated in PVL [6]. In rodents, NF-κB is rapidlyactivated following HI [7]. Pharmacological inhibition ofNF-κB after HI significantly reduced brain injury andlong-term motor and cognitive impairments [7–9].While these studies implicate NF-κB in the pathogenesisof PVL, the specific cell lineages responsible for NF-κB-mediated cell death remain unknown.Our data demonstrate that microglial activation by

NF-κB is a novel, cell-specific target for PVL pathogen-esis. In this study, we demonstrate that prophylactic se-lective inhibition of NF-κB in the microglia dramaticallyattenuates HI-induced white matter injury in a mousemodel of hypoxic ischemic injury of the immature brain.Inhibition of NF-κB in microglia protected mice fromhypomyelination and ventriculomegaly by decreasingproinflammatory mediators and sparing of myelin.

Materials and methodsAnimalsAll procedures were performed in accordance with theNIH Guidelines on the care and use of vertebrate ani-mals and approved by the Institutional Animal Care andUse Committee of the Feinstein Institutes for MedicalResearch and University of Arizona. Animals werehoused in a 12-h light/dark cycle in a virus/Ag-freefacility with controlled temperature and humidity andprovided with water and food ad libitum. Colony stimu-lating factor receptor 1 Cre IKKβflox/wt (CSF-1R-icre;IKKβflox/wt) were generated by breeding C57BL/6 CSF-1R-cre mice (Jackson Laboratories) to IKKβflox/flox mice(Jackson Laboratories). Cre specificity was confirmed bycrossing cre lines to C57BL/6 Rosa26-StopFlox-CAG-tdTomato (Jackson Laboratories) mice and assessed fortdTomato expression by immunohistochemistry. Equalnumber of male and female mice was used for thestudies.

Genotypes of used animalsGenotypes for all studied pups were determined byqualitative PCR using the following primers:

Gene Forward primer (5′-3′) Reverse primer (5′-3′)

IKKβ GTC ATT TCC ACA GCC CTGTGA

CCT TGT CCT ATA GAAGCA CAA

iCre+ CAGGGCCTTCTCCACACCAGC CTGGCTGTGAAGACCATC

Cre− GGACATGTTCAGGGATCGCCAGGCG

CGACGATGAAAGCATGTTTAGCTG

tdTomato+ CTG TTC CTG TAC GGC ATGG

GGC ATT AAA GCA GCGTAT CC

tdTomato− AAG GGA GCT GCA GTG GAGTA

CCG AAA ATC TGT GGGAAG TC

Hypoxia ischemia insultFor our study, we used our animal model of HI injury ofthe immature brain which was described in detail before[10]. In brief, under complete aseptic precautions,postnatal day 5 (P5) mouse pups were anaesthetizedwith isoflurane, midline neck incision was performed,and both carotid arteries were temporary ligated (using6.0 silk sutures double knot) for 10 min. After that,sutures were removed and the neck incision closed.Pups were allowed to recover for 30 min on thermalblanket, then were placed in hypoxia chamber (FiO2 8%)for 20 min after which they returned to their dams.During surgery and recovery and in the hypoxicchamber, pups were placed on thermal blanket and theirrectal temperature maintained at 36.5 °C measured by anultrathin rectal probe (BiosebLab, France). Shamcontrols at P5 were anaesthetized with isoflurane, andmidline neck incision was performed. Both carotid

Zaghloul et al. Journal of Neuroinflammation (2020) 17:365 Page 2 of 13

arteries were isolated but not ligated. Then, the neckincision was closed. During surgery and recovery, shammice were placed on thermal mattress and rectaltemperature was maintained at 36.5 °C.

Histopathological examBrain tissue was fixed on P15 in 4% paraformaldehydefor 24 h, processed, embedded in paraffin, andsubsequently cut into 6-μm-thick sections. Followingdeparaffinization, hematoxylin and eosin (H&E) stainingwas performed according to standard protocols. Typicalsections of hippocampus and cerebrum were made ineach group of animals.

ImmunohistochemistryAnimals were deeply anesthetized on P15 with alethal dose of xylazene/ketamine and perfusedtranscardially with normal saline, then 4%paraformaldehyde. Brains were sectioned coronally at6-μm-thick using a microtome. Sections were incu-bated for 2 h at room temperature in TBS+ 1%Triton-X + 10% donkey serum. Samples were incu-bated for 24 h at 4 °C with primary antibodies,followed by 2-h incubation at RT with secondaryantibodies. All images were captured on a Zeiss con-focal microscope (Carl Zeiss, Thornwood, NY, USA).The following primary antibodies were used to detectthe following markers: tomato lectin (Vector labora-tories, Burlingame, CA, USA), CD 68 (AbS Serotec {1:100} [11, 12], Raleigh, NC, USA), cleaved caspase 3(Cell Signaling Technology {1:50}, Danvers, MA,USA), GFAP (Abcam {1:500} Cambridge, MA, USA),Iba1 (Wako {1:400}, Richmond, VA, USA), CNPase(Abcam {1:200} Cambridge, MA, USA), NeuN (EDMMillipore {1:250} Billerica, MA, USA), Olig2 (SantaCruz Biotechnology{1:50} Dallas, TX, USA), and sec-ondary antibodies (Species specific Cy3 and FITC 1:125 Jackson Immunoresearch, Westgroove, PA, USA).N = 5 animals/group.

Immunostaining analysisDigital images obtained from Confocal software wereexported to ImageJ. The fluorescence intensity associatedwith each pixel was determined in sections 750 × 750 μmthat included 4 sections per animal and 5 animals pergroup. Excitation and acquisition parameters wereadjusted to fully eliminate pixel saturation, and all imageswere collected under identical settings. Cell counting wasperformed on 4 sections per animal (750 × 750 μm each)and 5 animals per group.

Inflammatory cytokines assayAssay of proinflammatory cytokines IL-1β, IL-6, andTNF-α was done on the periventricular brain area on

postnatal day 6 using Quantikine ELISA kits (R&DSys-tems) which was used according to the manufacturer’sinstructions. N = 8 animals/group.

MRI instrumentation and data acquisitionMRI data were obtained on P15 using 9.4/30 BioSpectSpectrometer (Bruker BioSpin Corp., Germany) equippedwith a 72-mm volume coil as a transmitter and a 4-channel mouse brain coil. High-resolution RARE T2-weighted images in axial and coronal plain were acquiredwith following parameters: TR/TE = 3524/36ms, FOV =1.5 cm, matrix = 256 × 256, 0.5 mm slice thickness,Nslices = 30, Navg = 4, and RARE Factor = 8.The animals were anesthetized on P15 using 1–2%

isoflurane with carbogen (95%O2+5%CO2). Theirrespiration and temperature were monitored during theentire course of the experiment using Small AnimalInstruments monitoring system (Small AnimalsInstruments, Inc., Stony Brook, NY, USA).

MRI data analysisThe axially acquired T2-weighted images (0.5 mm slicethickness) were used for data analysis. Mask of themouse brain and ventricles were obtained by manuallytracing area of interest. Coronally acquired images wereused for a conformation. Images were imported intoAnalyze 7.5 software (Biomedical Imaging Resource,Mayo Clinic, Rochester, MN) for manual and semi-automated volume rendering. We estimated brainvolumes for the 0.5-mm gap using Analyze. Total braintissue volume was defined as the total intracranialvolume minus all cerebrospinal fluid (CSF) spaces. N =10 animals/group.

Neurobehavioral testing and long-term outcomeForelimb and hindlimb grip strength were measuredusing a grip strength meter (Columbus Instruments) at60 days of age. Each session consisted of the average ofthree tests per animal. Rotarod device (ColumbusInstruments, Columbus, OH, USA) was used to measuremotor function and balance at P60. Each sessionconsisted of the average of three trials on the elevatedaccelerating rotarod beginning at 5 RPM, measuring thetime the mouse was able to remain on the rod. At P60,animals were tested in an open field analysis (San DiegoInstruments, San Diego, CA, USA). Animals were givenseveral minutes to adapt to the testing chamber beforethe beginning of testing. Open field data was digitallyrecorded for 30 min and subsequently analyzed byNoldus Ethovision tracking software [13]. Beam breakswere recorded in the x, y, and z planes and averagedacross groups. N = 25 animals/group.

Zaghloul et al. Journal of Neuroinflammation (2020) 17:365 Page 3 of 13

Statistical analysesFor all statistical tests, Graph Pad Prism 7 software (LaJolla) was used. Statistical analysis of mean differencesbetween groups was performed by using one-wayANOVA, followed by a Bonferroni post-hoc analysis. Allp values and n values are indicated in figure legends. Nosex difference was found in the studies.

ResultsCSF-1R-cre is expressed mainly in microglia in our mousemodelTo determine the relevance of NF-κB inhibition in mye-loid cells mainly microglia in PVL pathology, we crossedmice with conditional mutant of IKKb (IKKbf/f), whichhave exon 3 of the ikbkb (IKKb) gene flanked by loxPsites to a strain expressing cre recombinase driven bythe promoter for the gene c-fms, which encodes colonystimulating factor receptor 1 (CSF-1R). CSF-1R isexpressed throughout the mononuclear phagocytesystem of the mouse, including monocyte-derived mac-rophages, although mainly microglia express CSF-1R inthe postnatal mouse brain [14]. Since homozygous micedevoid of IKKb in myeloid cells display severe immunedysfunction, we evaluated heterozygous mice. Heterozy-gous deletion of IKKb in myeloid cells results in ap-proximately a 50% reduction in NF-κB activity in theCNS.To confirm that cre expression was restricted to Iba1-

expressing microglia in the brain, we crossed CSF-1R-cre mice to a Rosa26 line that expresses tdTomato (RFP)in all cre-expressing cells. We observed robust RFP ex-pression in Iba1+ microglia throughout the neonatalmouse brain including microglia near the lateral ventri-cles (Fig. 1a) and in the cortex (Fig. 1b). This proves thatcre expression is restricted to microglia.Since there is no activation of NF-κB without hypoxia

ischemia, sham room air (RA) Cre + and sham (RA) Cre– showed the same results in all histopathology and mo-lecular testing. Therefore, through the experiments, oneRA group was displayed.Neonate mice were subjected to hypoxia ischemia

insult on P5 by temporary bilateral carotid arteryligation followed by transfer to hypoxia chamber.Three mice groups were studied. RA or Sham group,

CSF-1R cre – IKKβflox/wt mice subjected to HI (HICre–), and CSF-1R cre + IKKβflox/wt mice subjected toHI (HI Cre +).

NF-κB inhibition in microglia leads to white mattervolume preservationMRI volumetric analysis at P15 (10 days post-HI), usingthe BioSpec 94/30 Imaging 9.4 T, detected enlarged ven-tricles in mice subjected to hypoxia ischemia (Fig. 2a).We observed a significant decrease of lateral ventricle

volume in HI Cre+ as compared to HI Cre− mice (P <0.05) (Fig. 2b). There was no significant difference intotal brain volume between the three groups (Fig. 2c).The ratio of ventricle volume to total brain volume wassignificantly decreased in the HI Cre+ in comparison tothat in the HI Cre− group (P < 0.05) (Fig. 2d). Less ven-triculomegaly indicates white matter volume preserva-tion in HI Cre+.

NF-κB inhibition in microglia decreases ventriculomegalyand neuronal cell death in HI modelThe HI Cre− group had markedly enlarged lateralventricles and third ventricle shown by hematoxylin andeosin staining, while HI Cre+ only had minimal increasein lateral and third ventricle size comparable to RAcontrols. Ventricle enlargement which is caused bywhite matter loss after HI was strikingly reduced in theHI Cre + group and resembled non-injured controls(Fig. 3). HI Cre+ show reduced neuronal damage in cor-tex and hippocampus compared to the HI Cre− group.Ependymal lining was thickened and had increased cel-lular mitosis, indicating injury in the HI Cre− groupcompared to the HI Cre+ group. These findings indicatethat prophylactic inhibition of NF-κB in the microgliadecreased neuronal damage and oligodendroglial cellloss (Fig. 3).

NF-κB inhibition in microglia rescues oligodendroglia inHI modelTo provide further evidence that decreasedventriculomegaly is due to myelin preservation, westained for 2′,3′-cyclic-nucleotide 3′-phosphodiesterase(CNPase). Levels of CNPase were significantly decreasedin the HI Cre− group as compared to those in the RAcontrol (Fig. 4a). There was only a mild decreased inCNPase in the HI Cre+ group indicatingoligodendrocyte differentiation and white matterpreservation (Fig. 4a). Similar results were shown usingOligodendrocyte transcription factor (Olig 2), anoligodendrocyte precursor (Fig. 4b). Thus, prophylacticinhibition of NF-κB in the microglia increase oligo-dendroglial precursors and promotes differentiation.

NF-κB modulates microglial activation to aninflammatory, neurotoxic phenotypeWe hypothesized that the sparing of white matter in HICre+ pups was due to decreasing the proinflammatorymicroglia response. Therefore, we evaluated HI Cre−and HI Cre+ pups for CD68, a marker for activatedproinflammatory microglia. Microglia in mice subjectedto HI exhibit an activated proinflammatory, CD68+phenotype. HI Cre+ mice displayed reduced CD68+microglia in the periventricular white matter comparedto the HI Cre− mice (Fig. 5a).

Zaghloul et al. Journal of Neuroinflammation (2020) 17:365 Page 4 of 13

NF-κB inhibition in microglia decreases proinflammatorymilieuAdditionally, proinflammatory mediators IL1β, IL6, andTNF-α, which are increased in human PVL, were alsosignificantly increased in HI Cre− mice as compared toHI Cre+ mice (P < 0.05) (Fig. 6). Collectively, these datademonstrate reducing NF-κB activity in microglia damp-ened proinflammatory microglial activation in a mousemodel of PVL.

NF-κB activation increases nitrative stress, causing injuryto the vulnerable oligodendrogliaFree radical injury to the developing oligodendroglialcells (OL) underlies the pathogenesis of PVL and thehypomyelination seen in long-term survivors. In humanPVL, free radical injury leads to increased lipid peroxida-tion and nitrotyrosine production. Nitrotyrosine is se-creted by microglia as shown by the co-localization ofnitrotyrosine with Iba-1. Nitrotyrosine expression is sig-nificantly increased in HI Cre− as compared to HI Cre+(P < 0.05) (Fig. 7). Pre-oligodendrocytes are the mostvulnerable cells to nitrative stress; therefore, theyundergo cell loss leading to ventriculomegaly. HI Cre+group exhibits reduction in nitrative stress as a result ofdecrease nitrotyrosine release from microglia. This leadsto a decrease in pre-oligodendroglial loss in the HI Cre+as compared to the HI Cre− group.

Improved long-term neurobehavioral outcome withmicroglial inhibition of NF-κBTo investigate if the improvement in histopathologicalfindings and oligodendroglial survival along with areduction in inflammation translate to improvement inneurodevelopmental outcome in HI Cre+ compared toHI Cre− group, neurobehavioral analysis was performedat P60. Rear grip strength was significantly higher in HICre+ compared to HI Cre− group mice (Fig. 8a). Frontgrip strength was also improved in HI Cre+ versus HICre− group (Fig. 8a). On rotarod, HI Cre− group had amore tendency to fall indicating impaired coordinationin HI Cre− as compared to HI Cre+ group mice(Fig. 8b). By analyzing open field, number of rears andbeam breaks were also significantly decreased, indicatingworsening locomotion in HI Cre− versus HI Cre+ groupmice (Fig. 8c). These long-term findings mimic humanPVL neurodevelopmental outcomes such as diplegia/paresis of the lower limbs, incoordination, and tendencyto fall along with reduced activity and attention.

DiscussionPeriventricular leukomalacia (PVL) is a majorneuropathologic brain injury and is the most commoncause of cerebral palsy (CP) in premature infants.Currently, there is no available treatment for thisdevastating injury. Microglia are one of the maininflammatory cells involved in the pathogenesis of PVL.

Fig. 1 CSF-1R-cre is selectively expressed in microglia in the CNS. CSF-1R is expressed in microglia. CSF-1R-cre mice was crossed to a Rosa26 linethat expresses tdTomato (RFP) confirms microglial Iba1 expression in all cre-expressing cells. a Robust RFP expression in Iba1+ microgliathroughout the neonatal mouse brain including microglia in the periventricular area. b RFP expression in Iba1 microglia in the cortex. This provesthat the cre expression is restricted to microglia

Zaghloul et al. Journal of Neuroinflammation (2020) 17:365 Page 5 of 13

Here, we show that NF-κb activation in microglia isplaying a major role in the pathogenesis of hypoxic is-chemic injury of the immature brain. Heterozygous in-hibition of NF-κB in microglia offers significantprotection after exposure to HI insult. These findingssuggest specific inhibition of microglial NF-κb could bea novel target in the prophylaxis and treatment of HI in-jury of the immature brain.Phenotypically, this HI mouse model mimics human PVL

in the form of hind limb paresis and incoordination. Itdisplays great similarity to premature infants with severePVL who exhibit diplegia, incoordination, and CP [10].Also, this model mimics human PVL histologically in theform of white matter loss leading to ventriculomegaly, witholigodendroglial cell loss and maturational arrest [10].During HI insult, an inflammatory response cascade is

activated, which leads to massive cell damage andnecrosis. Hypothermia or hyperthermia can alter this

inflammatory response and thus the degree of HI injury.In our study, temperature was measured during surgeryand in the hypoxia chamber and was maintained at36.5 °C in both HI groups. Pups were returned to theirdams, and no further measurements were obtained due totechnical difficulties given the small size of the pups (2–3g). Dams keep pups warm and keep them normothermic,but there is a potential that pups became hypothermiaafter HI which could potentially alter the degree of HIinjury. Recent human studies support contributory rolefor proinflammatory cytokines in the pathogenesis of PVL[15]. Incidence of PVL and CP in premature infants isincreased in the presence of maternal, placental, or fetalinfection [16–27]. Elevated levels of IL-6 in the cord blood[3, 28], elevated levels of IL-6 and IL-1β in amniotic fluid[4, 29], and elevated levels of all interferons and IL-1 andIL-6, among other cytokines, in neonatal blood have allbeen associated with increased incidence of PVL and CP

Fig. 2 NF-KB inhibition in microglia leads to white matter volume preservation. MRI volumetric analysis at postnatal day 15 comparing sham RA,HI IKKβflox/wt CSF-1R Cre− (HI Cre−), and HI IKKβflox/wt CSF-1R Cre+ (HI Cre+), using the BioSpec 94/30 imaging system is a 9.4 T horizontal boremagnet operates at 400 MHz and runs ParaVision™4.0 software. a Coronal (top) and axial (bottom) brain slices. b The two lateral ventricle volumesmeasured by paravision software. c Total brain volume includes brain volume + lateral ventricle volume measured by paravision software. d Theratio of both lateral ventricle volumes to total brain volume measured by paravision software. N = 10 animals/group. Scatter dot plot showingmean ± SE. indicates*P < 0.05. **P ≤ 0.01

Zaghloul et al. Journal of Neuroinflammation (2020) 17:365 Page 6 of 13

in premature infants [19, 30, 31]. In animal models, IL-1β,IL-6, MIP, IL-9, and TNF-α seem to play an importantrole [15, 32, 33]. Cytokines and bacterial products cancause direct injury to the developing OLs. It was shownthat TNF-α is toxic to OLs [11, 34–40]; others demon-strated that OLs show high toxicity by interferon-y. Ιmma-ture OLs in a culture are more vulnerable to thecytotoxicity of interferon-y than are mature OLs [6, 36,41, 42]. In addition, TNF-α potentiates this toxicity ofinterferon-y to developing OLs [39].Anti-inflammatory treatments may represent a

useful strategy in the treatment of PVL, whereclinical conditions would favor a post-insult treat-ment strategy. Minocycline, as a microglial in-activator, showed protective effect when adminis-tered following HI insult [5]. Attenuation of inflam-matory reaction by pharmacological inhibition ofNF-κB significantly reduced brain injury after HI in-sult as shown by improvement in long-term motorand cognitive functions [8, 9]. In our study, we ex-plored both the dynamic and cell-specific activationof NF-κB during HI insult and found that specificinhibition of NF-κB in microglia in neonate mice

exposed to HI insult leads to decrease microglia acti-vation. A decrease in microglial activation led to lessoligodendroglial destruction (Fig. 4).It is well established in animal models that ischemia

reperfusion is accompanied rapidly by activation ofmicroglia, secretion of cytokines, and mobilization,adhesion, and migration of macrophages andinflammatory cells [43, 44]. While microglia are themain cells in the brain that express CSF-1R, mono-cytes and macrophages also express CSF-1R and canbe seen in the brain in HI due to increased bloodbrain barrier permeability. Therefore, it is likely thatpart of this neuroprotective effect is due to inhibitionof NF-κB in infiltrating monocytes and macrophages.Whether induced by infection or ischemia, these in-flammatory responses could be detrimental to devel-oping OLs [35, 45, 46]. Vasoactive effects of certaincytokines and nitrogen species released, as part of theinflammatory cascade, can impact cerebrovascularregulation, impair perfusion, and thereby increase therisk for ischemic injury [47, 48]. In HI Cre+ mice,there was a significant reduction of inflammatory

Fig. 3 NF-κB inhibition in microglia decreases ventriculomegaly and neuronal cell death in PVL model. Representative H&E coronal brain sectionsof lateral ventricle, ependymal lining, cortex, and hippocampus of the studied groups, sham RA, HI IKKβflox/wt CSF-1R Cre− (HI Cre−), and HIIKKβflox/wt CSF-1R Cre+ (HI Cre+) at postnatal day 15. Blue arrows point to neuronal cell death. Scale bar 10 μm except for lateral ventricle panel is40 μm. N = 12 animals/group

Zaghloul et al. Journal of Neuroinflammation (2020) 17:365 Page 7 of 13

cytokines within 24 h post-HI insult as compared toHI Cre− mice (Fig. 6).Diffuse oligodendroglial (OL) injury in PVL is

related to moderate ischemia. Early differentiating OLor pre-OL is vulnerable to free radical attack, whereasthe mature OL is resistant [49, 50]. Increase accumu-lation of free radicals in OL precursors, with limitedand deficient antioxidants defense, leads to hydrogenperoxide accumulation, which produces the deadly hy-droxyl radical [51–53]. In human PVL, free radical in-jury is supported by evidence of oxidative andnitrative stress with markers to lipid peroxidation andnitrotyrosine [2]. Free radicals are both a cause and aresult of inflammation. Reperfusion of ischemic tis-sues is associated with microvascular injury, particu-larly due to increased permeability of capillaries andarterioles, which leads to an increase of diffusion andfluid filtration across the tissues. These activated in-jured endothelial cells produce more reactive oxygenand nitrogen species which trigger more inflammatory

response. NF-κB can also induce nitric oxide synthasein glial cells resulting in the production of nitricoxide (NO) and related neurotoxic reactive oxygenspecies [54]. NO itself can induce pre-OL damage bytwo mechanisms: one involving the direct effect of ni-tric oxide on pre-OL mitochondrial integrity andfunction, and the other involving an activation ofmicroglia and subsequent release of reactive nitrogenspecies [12]. Activated microglia release NO and ni-trogen species, which mediate neurotoxicity in severalneurodegenerative diseases and in HI insults [55, 56].In our model, inhibition of NF-κB in microglia, notonly attenuates the inflammatory response elicited bycell injury and free radical accumulation, but alsomodulates NO release. Our results showed a signifi-cant reduction of activated microglia in HI Cre+mice, as one of the main sources of released NO andnitrotyrosine (Figs. 5 and 7).In our HI model, long-term neurobehavioral evalu-

ation showed a significant decrease in rear grip

Fig. 4 NF-KB inhibition in microglia rescues oligodendroglia in HI model. a CNPase in red (early oligodendroglia differentiation marker) in theperiventricular area of RA, HI IKKβflox/wt CSF-1R Cre− (HI Cre−), and HI IKKβflox/wt CSF-1R Cre+ (HI Cre+) at postnatal day 15. Quantification ofCNPase intensity as fold change where sham = 1. b Olig 2 in red (early oligodendroglia transcription factor) in the periventricular area of RA, HIIKKβflox/wt CSF-1R Cre− (HI Cre−), and HI IKKβflox/wt CSF-1R Cre+ (HI Cre+) at postnatal day 15. Quantification of Olig2-positive cells per section.Scale bar = 100 μm N = 5 animals/ group. Four sections/animal. Scatter dot plot showing mean ± SE. *P < 0.05. **P ≤ 0.01. ***P ≤ 0.001, ****P≤ 0.0001

Zaghloul et al. Journal of Neuroinflammation (2020) 17:365 Page 8 of 13

strength, overall locomotion, and incoordination mim-icking human PVL who present with diplegia and in-coordination (Fig. 8). These findings can be explained byperiventricular white matter loss as evident in our MRIstudies showing significantly increased ventriculomegalyin HI group assessed by ventricle volume as well as by cal-culating the ventricle/brain volume ratio (Fig. 2). The ven-tricle volume was statistically increased in HI Cre− groupat 2.5 times the volume of the HI Cre+ group. The HI Cre

− group had slightly smaller total brain volume than theHI Cre + group but this was not statistically significant.Ventriculomegaly and white matter loss are also shown inour histopathological studies (Fig. 3). Severity of motorimpairment and cognitive impairment were found to beclosely associated with the lateral ventricular volumes[57]. Reduction of the periventricular white matter andthe degree of lateral ventricle expansion are the maincauses of dysfunctions and damage of vision and hearing

Fig. 5 NF-κB modulates microglial activation to an inflammatory, neurotoxic phenotype. a Immunostaining for all microglia (Iba1) in green (top)and CD 68 (marker of activated microglia M1 only) in red (middle). Co-localization in yellow indicating amount of activated microglia at postnatalday 15. * indicates lateral ventricle. Quantification of CD68/Iba1 per section. Scale bar = 100 μm N = 5 animals/group and 4 sections/animal.Scatter dot plot showing mean ± SE. ***P ≤ 0.001, ****P ≤ 0.0001

Fig. 6 NF-KB inhibition in microglia decreases proinflammatory milieu. ELISA assay of proinflammatory cytokines IL-1β, IL 6, and TNF-α from brainhomogenates at postnatal day 6. These cytokines are known to peak in both human and animal PVL models HI. N = 8 animals/ group. Scatterdot plot showing mean ± SE. **P ≤ 0.01. ***P ≤ 0.001, ****P ≤ 0.0001

Zaghloul et al. Journal of Neuroinflammation (2020) 17:365 Page 9 of 13

as well as intellectual impairment and CP in children withPVL [58, 59]. Specific inhibition of NF-κB in microglia inneonate mice exposed to HI insult led to a significant at-tenuation of ventriculomegaly as shown by measuringboth ventricle volumes. These MRI Findings, in additionto the decreased neuronal damage and oligodendroglialloss, shown in our histopathological studies, explain thesignificant improvement in neurodevelopmental long-term outcome among HI Cre+ group.

NF-κB is a key transcription factor involved in theregulation of cytokine production [60]. NF-κB canmodulate the expression of apoptosis-promoting cyto-kines such as TNF-α and FAS ligand (FASL) [61]. Inmodels of ischemia, NF-κB activation appears to con-tribute to brain damage and mice lacking the p50subunit of NF-κB demonstrate decreased infarct vol-umes [62, 63]. Many studies highlight the pivotal roleof activation of NF-κB in glial cells that lead to

Fig. 7 NF-κB activation increases nitrative stress, causing injury to the vulnerable oligodendroglia. a Immunohistochemistry of coronal sections ofthe periventricular brain area of postnatal day 15 pups. Top panel shows Nitrotyrosine (marker of nitrative stress) in red. Middle panel shows Iba1(microglial marker) in green. Lower panel shows Co-localization in yellow nitrotyrosine secreted by microglia. Quantification of nitrotyrosine/Iba1per section. Scale bar = 200 μm. N = 5 animals/group and 4 sections/animal. Scatter dot plot showing mean ± SE. *P < 0.05, ****P < 0.0001. bHigher magnification in the HI IKKβflox/wt CSF-1R Cre− (HI Cre−) group showing the co-localization of nitrotyrosine to Iba1 indicating thatmicroglia is source of the nitrative stress (nitortyrosine)

Fig. 8 Improved long-term neurobehavioral outcome with microglial inhibition of NF-KB. a Testing for rear and front grip strength at postnatalday 60 (P60). b Testing of latency to fall in seconds by rotarod for coordination assessment at P60. c Number of beam breaks and number ofrears in the open field test as an assessment of motor function at P60. N = 25 animals/group. Scatter dot plot showing mean ± SE. **P ≤ 0.01,***P ≤ 0.001, ****P ≤ 0.0001

Zaghloul et al. Journal of Neuroinflammation (2020) 17:365 Page 10 of 13

production of neurotoxins [64]. In a trial to attenuatethe proinflammatory reaction of astroglia mediated byNF-κB expression, transgenic models with relativelyinhibited NF-κB expression were used and showed asignificant reduction in white matter injury in diseasemodels of autoimmune encephalomyelitis and ische-mic stroke [8]. Other studies used certain small cell-penetrating peptides (CPPs) to inhibit NF-κB activa-tion in animal models of different disorders, e.g., dia-betes mellitus type 1 [65], inflammation (acute andchronic) [12, 66–68], and cancer [69, 70]. Most ofCCPs studies (in vitro and in vivo studies) showedsignificantly attenuated inflammatory infiltration, de-creased cell necrosis, and degeneration with resultantamelioration of disease severity [71]. Our study aswell as others suggests that focusing on NF-κB activa-tion is a realistic target for the development of thera-peutic approaches.

ConclusionNF-κB regulates a large number of essential cellularactivities and is essential for many developmentalaspects. This precludes “blanket” inhibition of NF-κBas a clinical intervention. NF-κB activation in glutami-nergic neurons is involved in learning and memory,dendritic arborization, and axonal outgrowth [72].When targeting NF-κB in the brain, cell-type specifi-city is required so the functional benefits are achievedby dampening microglial activation and inflammation.Our study highlights the fact that NF-κB activation inmicroglia is a major contributor of white matter in-jury in the premature brain. Using a specific inhibitorof microglial NF-κB can be a new prophylactic/thera-peutic approach for hypoxic ischemic brain injury.

AbbreviationsPVL: Periventricular lekomalacia; CP: Cerebral palsy; HI: Hypoxia ischemia; CSF-1R Cre +: Colony stimulating factor receptor 1-Cre positive; RFP: Redfluorescent protein; CNPase: Cyclic-nucleotide 3′-phosphodiesterase; Olig2: Oligodendrocyte transcription factor; IL-1β: Interleukin 1β; IL-6: Interleukin6; TNF-α: Tumor necrosis factor-α; OL: Oligodendroglial cells; NO: Nitric oxide;FASL: FAS ligand; CPPs: Small cell-penetrating peptides

AcknowledgementsNot applicable.

Authors’ contributionsNZ, DK, MB, NN, and MNA conceived the project and designed theexperiments. NZ wrote the manuscript. NZ and MNA performed all in vivoexperiments. NZ and NN oversaw all breeding programs, as well as managedthe transgenic animal colony and genotyping. NZ and DK performed theimmunostaining and all statistical analyses. NN performed the cytokineprotein expression. NZ performed all the neurobehavioral testing. DK, MB,and MNA revised the manuscript. The authors read and approved the finalmanuscript.

FundingNot applicable.

Availability of data and materialsThe datasets used and/or analyzed during the current study are availablefrom the corresponding author on reasonable request.

Ethics approvalAll mice procedures were performed in accordance with the NIH Guidelineson the care and use of vertebrate animals and approved by the InstitutionalAnimal Care and Use Committee of the Feinstein Institutes for MedicalResearch and University of Arizona.

Consent for publicationNot applicable.

Competing interestsThe authors declare that they have no competing interests.

Author details1Department of Pediatrics, Division of Neonatology, University of Arizona,1501 N. Campbell Avenue, Tucson, AZ, USA. 2Department of Pediatrics,Division of Neonatology, Feinstein Institute for Medical Research, Manhasset,NY, USA.

Received: 19 August 2020 Accepted: 10 November 2020

References1. Walton M, Connor B, Lawlor P, Young D, Sirimanne E, Gluckman P,

et al. Neuronal death and survival in two models of hypoxic-ischemicbrain damage. Brain Res Brain Res Rev. 1999;29(2-3):137–68.

2. Haynes RL, Folkerth RD, Keefe RJ, Sung I, Swzeda LI, Rosenberg PA, et al.Nitrosative and oxidative injury to premyelinating oligodendrocytes inperiventricular leukomalacia. J Neuropathol Exp Neurol. 2003;62(5):441–50.

3. Billiards SS, Haynes RL, Folkerth RD, Trachtenberg FL, Liu LG, Volpe JJ, et al.Development of microglia in the cerebral white matter of the human fetusand infant. J Comp Neurol. 2006;497(2):199–208.

4. Rezaie P, Dean A. Periventricular leukomalacia, inflammation and white matterlesions within the developing nervous system. Neuropathology. 2002;22(3):106–32.

5. Lechpammer M, Manning SM, Samonte F, Nelligan J, Sabo E, Talos DM,et al. Minocycline treatment following hypoxic/ischaemic injury attenuateswhite matter injury in a rodent model of periventricular leucomalacia.Neuropathol Appl Neurobiol. 2008;34(4):379–93.

6. Volpe JJ. Neurobiology of periventricular leukomalacia in the prematureinfant. Pediatr Res. 2001;50(5):553–62.

7. van der Kooij MA, Nijboer CH, Ohl F, Groenendaal F, Heijnen CJ, van Bel F, et al. NF-kappaB inhibition after neonatal cerebral hypoxia-ischemia improves long-termmotor and cognitive outcome in rats. Neurobiol Dis. 2010;38(2):266–72.

8. Brambilla R, Persaud T, Hu X, Karmally S, Shestopalov VI, Dvoriantchikova G,et al. Transgenic inhibition of astroglial NF-kappa B improves functionaloutcome in experimental autoimmune encephalomyelitis by suppressingchronic central nervous system inflammation. J Immunol. 2009;182(5):2628–40.

9. Dvoriantchikova G, Barakat D, Brambilla R, Agudelo C, Hernandez E, BetheaJR, et al. Inactivation of astroglial NF-kappa B promotes survival of retinalneurons following ischemic injury. Eur J Neurosci. 2009;30(2):175–85.

10. Zaghloul N, Patel H, Ahmed MN. A model of periventricular leukomalacia (PVL)in neonate mice with histopathological and neurodevelopmental outcomesmimicking human PVL in neonates. PLoS One. 2017;12(4):e0175438.

11. Louis JC, Magal E, Takayama S, Varon S. CNTF protection ofoligodendrocytes against natural and tumor necrosis factor-induced death.Science. 1993;259(5095):689–92.

12. Zhao J, Zhang L, Mu X, Doebelin C, Nguyen W, Wallace C, et al.Development of novel NEMO-binding domain mimetics for inhibiting IKK/NF-κB activation. PLoS Biol. 2018;16(6):e2004663.

13. Hartman R, Lekic T, Rojas H, Tang J, Zhang JH. Assessing functional outcomesfollowing intracerebral hemorrhage in rats. Brain Res. 2009;1280:148–57.

14. Erblich B, Zhu L, Etgen AM, Dobrenis K, Pollard JW. Absence of colonystimulation factor-1 receptor results in loss of microglia, disrupted braindevelopment and olfactory deficits. PLoS One. 2011;6(10):e26317.

15. Hamada Y, Hayakawa T, Hattori H, Mikawa H. Inhibitor of nitric oxidesynthesis reduces hypoxic-ischemic brain damage in the neonatal rat.Pediatr Res. 1994;35(1):10–4.

Zaghloul et al. Journal of Neuroinflammation (2020) 17:365 Page 11 of 13

16. Perlman JM, Risser R, Broyles RS. Bilateral cystic periventricular leukomalacia inthe premature infant: associated risk factors. Pediatrics. 1996;97(6 Pt 1):822–7.

17. Zupan V, Gonzalez P, Lacaze-Masmonteil T, Boithias C, d'Allest AM, DehanM, et al. Periventricular leukomalacia: risk factors revisited. Dev Med ChildNeurol. 1996;38(12):1061–7.

18. Grether JK, Nelson KB, Emery ES 3rd, Cummins SK. Prenatal and perinatal factorsand cerebral palsy in very low birth weight infants. J Pediatr. 1996;128(3):407–14.

19. Nelson KB, Dambrosia JM, Grether JK, Phillips TM. Neonatal cytokines andcoagulation factors in children with cerebral palsy. Ann Neurol. 1998;44(4):665–75.

20. Baud O, Ville Y, Zupan V, Boithias C, Lacaze-Masmonteil T, Gabilan JC, et al.Are neonatal brain lesions due to intrauterine infection related to mode ofdelivery? Br J Obstet Gynaecol. 1998;105(1):121–4.

21. Dammann O, Allred EN, Veelken N. Increased risk of spastic diplegia amongvery low birth weight children after preterm labor or prelabor rupture ofmembranes. J Pediatr. 1998;132(3 Pt 1):531–5.

22. Hansen A, Leviton A. Labor and delivery characteristics and risks ofcranial ultrasonographic abnormalities among very-low-birth-weightinfants. The Developmental Epidemiology Network Investigators. Am JObstet Gynecol. 1999;181(4):997–1006.

23. Redline RW, Wilson-Costello D, Borawski E, Fanaroff AA, Hack M. Therelationship between placental and other perinatal risk factors for neurologicimpairment in very low birth weight children. Pediatr Res. 2000;47(6):721–6.

24. Resch B, Vollaard E, Maurer U, Haas J, Rosegger H, Müller W. Risk factors anddeterminants of neurodevelopmental outcome in cystic periventricularleucomalacia. Eur J Pediatr. 2000;159(9):663–70.

25. Wu YW, Colford JM Jr. Chorioamnionitis as a risk factor for cerebral palsy: ameta-analysis. Jama. 2000;284(11):1417–24.

26. Yoon BH, Romero R, Park JS, Kim CJ, Kim SH, Choi JH, et al. Fetal exposureto an intra-amniotic inflammation and the development of cerebral palsy atthe age of three years. Am J Obstet Gynecol. 2000;182(3):675–81.

27. De Felice C, Toti P, Laurini RN, Stumpo M, Picciolini E, Todros T, et al. Earlyneonatal brain injury in histologic chorioamnionitis. J Pediatr. 2001;138(1):101–4.

28. Yoon BH, Romero R, Yang SH, Jun JK, Kim IO, Choi JH, et al. Interleukin-6concentrations in umbilical cord plasma are elevated in neonates withwhite matter lesions associated with periventricular leukomalacia. Am JObstet Gynecol. 1996;174(5):1433–40.

29. Yoon BH, Jun JK, Romero R, Park KH, Gomez R, Choi JH, et al. Amniotic fluidinflammatory cytokines (interleukin-6, interleukin-1beta, and tumor necrosisfactor-alpha), neonatal brain white matter lesions, and cerebral palsy. Am JObstet Gynecol. 1997;177(1):19–26.

30. Grether JK, Nelson KB. Maternal infection and cerebral palsy in infants ofnormal birth weight. JAMA. 1997;278(3):207–11.

31. Grether JK, Nelson KB, Dambrosia JM, Phillips TM. Interferons andcerebral palsy. J Pediatr. 1999;134(3):324–32.

32. Yoon BH, Kim CJ, Romero R, Jun JK, Park KH, Choi ST, et al. Experimentallyinduced intrauterine infection causes fetal brain white matter lesions inrabbits. Am J Obstet Gynecol. 1997;177(4):797–802.

33. Debillon T, Gras-Leguen C, Verielle V, Winer N, Caillon J, Roze JC, et al.Intrauterine infection induces programmed cell death in rabbitperiventricular white matter. Pediatr Res. 2000;47(6):736–42.

34. Selmaj K, Raine CS, Farooq M, Norton WT, Brosnan CF. Cytokinecytotoxicity against oligodendrocytes. Apoptosis induced bylymphotoxin. J Immunol. 1991;147(5):1522–9.

35. Mayer M, Noble M. N-acetyl-L-cysteine is a pluripotent protector againstcell death and enhancer of trophic factor-mediated cell survival in vitro.Proc Natl Acad Sci U S A. 1994;91(16):7496–500.

36. Vartanian T, Li Y, Zhao M, Stefansson K. Interferon-gamma-inducedoligodendrocyte cell death: implications for the pathogenesis of multiplesclerosis. Mol Med. 1995;1(7):732–43.

37. Agresti C, D'Urso D, Levi G. Reversible inhibitory effects of interferon-gamma and tumour necrosis factor-alpha on oligodendroglial lineage cellproliferation and differentiation in vitro. Eur J Neurosci. 1996;8(6):1106–16.

38. Merrill JE, Ignarro LJ, Sherman MP, Melinek J, Lane TE. Microglial cell cytotoxicity ofoligodendrocytes is mediated through nitric oxide. J Immunol. 1993;151(4):2132–41.

39. Andrews T, Zhang P, Bhat NR. TNFalpha potentiates IFNgamma-induced celldeath in oligodendrocyte progenitors. J Neurosci Res. 1998;54(5):574–83.

40. Burgmaier G, Schönrock LM, Kuhlmann T, Richter-Landsberg C, BrückW. Association of increased bcl-2 expression with rescue from tumornecrosis factor-alpha-induced cell death in the oligodendrocyte cellline OLN-93. J Neurochem. 2000;75(6):2270–6.

41. Popko B, Baerwald KD. Oligodendroglial response to the immune cytokineinterferon gamma. Neurochem Res. 1999;24(2):331–8.

42. Baerwald KD, Popko B. Developing and mature oligodendrocytesrespond differently to the immune cytokine interferon-gamma. JNeurosci Res. 1998;52(2):230–9.

43. Dommergues MA, Patkai J, Renauld JC, Evrard P, Gressens P.Proinflammatory cytokines and interleukin-9 exacerbate excitotoxic lesionsof the newborn murine neopallium. Ann Neurol. 2000;47(1):54–63.

44. Tahraoui SL, Marret S, Bodenant C, Leroux P, Dommergues MA, Evrard P,et al. Central role of microglia in neonatal excitotoxic lesions of the murineperiventricular white matter. Brain Pathol. 2001;11(1):56–71.

45. Gehrmann J, Banati RB, Wiessner C, Hossmann KA, Kreutzberg GW. Reactivemicroglia in cerebral ischaemia: an early mediator of tissue damage?Neuropathol Appl Neurobiol. 1995;21(4):277–89.

46. Wilde GJ, Pringle AK, Sundstrom LE, Mann DA, Iannotti F. Attenuation andaugmentation of ischaemia-related neuronal death by tumour necrosisfactor-alpha in vitro. Eur J Neurosci. 2000;12(11):3863–70.

47. Brian JE Jr, Faraci FM. Tumor necrosis factor-alpha-induced dilatation ofcerebral arterioles. Stroke. 1998;29(2):509–15.

48. Haynes RL, Folkerth RD, Trachtenberg FL, Volpe JJ, Kinney HC. Nitrosativestress and inducible nitric oxide synthase expression in periventricularleukomalacia. Acta Neuropathol. 2009;118(3):391–9.

49. Back SA, Gan X, Li Y, Rosenberg PA, Volpe JJ. Maturation-dependentvulnerability of oligodendrocytes to oxidative stress-induced death causedby glutathione depletion. J Neurosci. 1998;18(16):6241–53.

50. Ferriero DM. Oxidant mechanisms in neonatal hypoxia-ischemia. DevNeurosci. 2001;23(3):198–202.

51. Fullerton HJ, Ditelberg JS, Chen SF, Sarco DP, Chan PH, Epstein CJ, et al.Copper/zinc superoxide dismutase transgenic brain accumulates hydrogenperoxide after perinatal hypoxia ischemia. Ann Neurol. 1998;44(3):357–64.

52. Groenendaal F, Shadid M, McGowan JE, Mishra OP, van Bel F. Effects ofdeferoxamine, a chelator of free iron, on NA(+), K(+)-ATPase activity ofcortical brain cell membrane during early reperfusion after hypoxia-ischemia in newborn lambs. Pediatr Res. 2000;48(4):560–4.

53. Thorburne SK. Low glutathione and high iron govern the susceptibility ofoligodendroglial precursors to oxidative stress. In: Juurlink BHJ, editor; 1996.

54. Akama KT, Albanese C, Pestell RG, Van Eldik LJ. Amyloid beta-peptidestimulates nitric oxide production in astrocytes through an NFkappaB-dependent mechanism. Proc Natl Acad Sci U S A. 1998;95(10):5795–800.

55. Lai AY, Todd KG. Hypoxia-activated microglial mediators of neuronal survivalare differentially regulated by tetracyclines. Glia. 2006;53(8):809–16.

56. Dommergues MA, Plaisant F, Verney C, Gressens P. Early microglialactivation following neonatal excitotoxic brain damage in mice: a potentialtarget for neuroprotection. Neuroscience. 2003;121(3):619–28.

57. Melhem ER, Hoon AH Jr, Ferrucci JT Jr, Quinn CB, Reinhardt EM, DemetridesSW, et al. Periventricular leukomalacia: relationship between lateralventricular volume on brain MR images and severity of cognitive and motorimpairment. Radiology. 2000;214(1):199–204.

58. Fedrizzi E, Inverno M, Bruzzone MG, Botteon G, Saletti V, Farinotti M. MRIfeatures of cerebral lesions and cognitive functions in preterm spasticdiplegic children. Pediatr Neurol. 1996;15(3):207–12.

59. Andiman SE, Haynes RL, Trachtenberg FL, Billiards SS, Folkerth RD, Volpe JJ,et al. The cerebral cortex overlying periventricular leukomalacia: analysis ofpyramidal neurons. Brain Pathol. 2010;20(4):803–14.

60. Baeuerle PA, Baltimore D. NF-kappa B: ten years after. Cell. 1996;87(1):13–20.61. Liu F, Bardhan K, Yang D, Thangaraju M, Ganapathy V, Waller JL, et al. NF-κB

directly regulates Fas transcription to modulate Fas-mediated apoptosis andtumor suppression. J Biol Chem. 2012;287(30):25530–40.

62. Schneider A, Martin-Villalba A, Weih F, Vogel J, Wirth T, SchwaningerM. NF-kappaB is activated and promotes cell death in focal cerebralischemia. Nat Med. 1999;5(5):554–9.

63. Nurmi A, Lindsberg PJ, Koistinaho M, Zhang W, Juettler E, Karjalainen-Lindsberg ML, et al. Nuclear factor-kappaB contributes to infarction afterpermanent focal ischemia. Stroke. 2004;35(4):987–91.

64. Mattson MP, Meffert MK. Roles for NF-kappaB in nerve cell survival,plasticity, and disease. Cell Death Differ. 2006;13(5):852–60.

65. Guardado Mendoza R, Perego C, Finzi G, La Rosa S, Capella C, Jimenez-CejaLM, et al. Delta cell death in the islet of Langerhans and the progressionfrom normal glucose tolerance to type 2 diabetes in non-human primates(baboon, Papio hamadryas). Diabetologia. 2015;58(8):1814–26.

Zaghloul et al. Journal of Neuroinflammation (2020) 17:365 Page 12 of 13

66. Paria BC, Bair AM, Xue J, Yu Y, Malik AB, Tiruppathi C. Ca2+ influxinduced by protease-activated receptor-1 activates a feed-forwardmechanism of TRPC1 expression via nuclear factor-kappaB activation inendothelial cells. J Biol Chem. 2006;281(30):20715–27.

67. Paria BC, Malik AB, Kwiatek AM, Rahman A, May MJ, Ghosh S, et al. Tumornecrosis factor-alpha induces nuclear factor-kappaB-dependent TRPC1expression in endothelial cells. J Biol Chem. 2003;278(39):37195–203.

68. Fukushima H, Jimi E, Okamoto F, Motokawa W, Okabe K. IL-1-inducedreceptor activator of NF-kappa B ligand in human periodontal ligamentcells involves ERK-dependent PGE2 production. Bone. 2005;36(2):267–75.

69. Huang TT, Wuerzberger-Davis SM, Wu ZH, Miyamoto S. Sequentialmodification of NEMO/IKKgamma by SUMO-1 and ubiquitin mediates NF-kappaB activation by genotoxic stress. Cell. 2003;115(5):565–76.

70. Ashikawa K, Shishodia S, Fokt I, Priebe W, Aggarwal BB. Evidence thatactivation of nuclear factor-kappaB is essential for the cytotoxic effects ofdoxorubicin and its analogues. Biochem Pharmacol. 2004;67(2):353–64.

71. Orange JS, May MJ. Cell penetrating peptide inhibitors of nuclear factor-kappa B. Cell Mol Life Sci. 2008;65(22):3564–91.

72. Kaltschmidt B, Ndiaye D, Korte M, Pothion S, Arbibe L, Prullage M, et al. NF-kappaB regulates spatial memory formation and synaptic plasticity throughprotein kinase A/CREB signaling. Mol Cell Biol. 2006;26(8):2936–46.

Publisher’s NoteSpringer Nature remains neutral with regard to jurisdictional claims inpublished maps and institutional affiliations.

Zaghloul et al. Journal of Neuroinflammation (2020) 17:365 Page 13 of 13